datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
safe-alignment-dynamic
safe-alignment-dynamic
Training prompts for score-conditioned SFT / RL and separate reward-model pair sets; nothing here is scored.
sft-prompts/train and rl-prompts/train: the same prompt pool, deduplicated across sources with responses
merged and HH/PKU test prompts removed. rl-prompts additionally marks selection=pku_label_conflict where PKU's
better and safer labels disagree with opposite safety flags; preference_pairs indexes those responses.
This is an annotation, not a… See the full description on the dataset page: https://huggingface.co/datasets/RLLab/safe-alignment-dynamic.MRRL-Mixed
MRRL-Mixed base
base/train is the shared, unbalanced prompt pool for model-specific sampling and calibration.
Sources
agentica-org/DeepScaleR-Preview-Dataset, revision b6ae8c60f5c1f2b594e2140b91c49c9ad0949e29.
PRIME-RL/Eurus-2-RL-Data, revision 9776b13264b5aaa0b16495fcf086a0a8d86fd655.
allenai/Dolci-RL-Zero-Code-7B, revision 054c0a1f9c5fd52566ad9124aa2e702924803cd3.
open-r1/verifiable-coding-problems-python_decontaminated-tested, revision… See the full description on the dataset page: https://huggingface.co/datasets/RLLab/MRRL-Mixed.eval-set
AIME sources
The aime24 and aime25 transcriptions are from MathArena by Jasper Dekoninck et al. Their CC BY-NC-SA 4.0 license applies to these two configs. The other configs retain their respective upstream licenses.
Pinned sources:
MathArena/aime_2024_I
MathArena/aime_2024_II
MathArena/aime_2025
Previous AIME mirrors used by AetherEval:
aime24: HuggingFaceH4/aime_2024.
aime25: yentinglin/aime_2025.
The pre-migration eval-set configs matched those AetherEval copies in all 30… See the full description on the dataset page: https://huggingface.co/datasets/RLLab/eval-set.nut-bolt-dataset
Hugging Face 모델 & 데이터셋 사용 가이드
이 페이지는 Hugging Face Hub에 올려진 YOLO26 모델과 데이터셋을 다운로드해서 사용하는 방법을 설명합니다.
모델 다운로드 및 추론
설치 및 로그인
pip install huggingface_hub ultralytics
huggingface-cli login # (또는 HUGGINGFACE_HUB_TOKEN 환경변수 설정)
모델 사용 - 기본 예시
from huggingface_hub import snapshot_download
from ultralytics import YOLO
model_dir = snapshot_download(
repo_id="your-id/nut-volt-yolo26m",
repo_type="model",
)
model =… See the full description on the dataset page: https://huggingface.co/datasets/rllab-postech/nut-bolt-dataset.OpenR1-Math-220k-FilteredRaR-Medicine-GroupedRaR-Science-Groupedmath-rlOpenR1-Math-220K-Filtered-DPOallenai-Dolci-Instruct-DPO-Filtered-GenerationsOpenR1-Math-220k-Filtered-Generationsallenai-Dolci-Instruct-DPO-Length-FilteredMTMR
MTMR (Multi-Task Multi-Reward) base
base/train is the shared, unbalanced prompt pool for model-specific sampling and calibration.
Sources
agentica-org/DeepScaleR-Preview-Dataset, revision b6ae8c60f5c1f2b594e2140b91c49c9ad0949e29.
open-r1/DAPO-Math-17k-Processed, revision 31dd309567e3da778038cc87d868b6097a3ccf68.
PRIME-RL/Eurus-2-RL-Data, revision 9776b13264b5aaa0b16495fcf086a0a8d86fd655.
allenai/Dolci-RL-Zero-Code-7B, revision… See the full description on the dataset page: https://huggingface.co/datasets/RLLab/MTMR.